Minimum Viable Ethical Product.
If ethics are so complex when we talk about AI, then we must prepare to build minimum ethical products (MEPs).
A minimum ethical product should have AI control elements, so as not to lose all the functionality of the product when it is paused due to some situation.
The idea is to comply with ethical standards and principles first and then meet business needs.
Implementing minimal ethical products also involves careful monitoring of how those products impact users.
We should ask questions like,
- What is the goal you are trying to achieve by creating this AI?
- Is it necessary to use AI to achieve this goal?
- Can you use other non-AI methods to achieve the same goal with lower risk? and finally,
- Is AI likely to be effective in helping us achieve this goal?
Related Posts
Alignment Values and Objectives.
